Information Note 97/3

April 1997

by E Ravanello

Audio and data services are carried in the Main Service Channel (MSC) of the Eureka 147 multiplex. This channel supports a gross bit rate of 2304 kbits/sec. However, the net bit rate (i.e. the actual capacity available for use) depends on the protection level applied to services. For audio-only services the net capacity of the ensemble varies between 783 (highest protection) and 1728 kbits/sec (lowest protection). The corresponding range for data-only services is 576 and 1728 kbits/sec. At a median protection level the available net capacity for both audio and data services is 1152 kbits/sec.

Within the MSC each audio or data service is carried in a sub-channel. Up to 63 sub-channels can be supported, each of which is treated individually as far as error protection is concerned.

2.1 Protection Level

In the case of audio services, five protection levels (1 to 5) have been specified in order to cater for a variety of applications. Level 5 affords the lowest protection and is designed for cable systems. It allows a high number of programme services, but does not have the strong error protection necessary for operation in multipath environments. Protection Level 3 is better suited to mobile operation.

To allow more flexibility in accommodating sub-channels, Protection Levels 4 and 2 have also been introduced with somewhat weaker and stronger performance than Protection Level 3 (respectively). Protection Level 1 is suited to applications with a very high sensitivity to transmission errors while Protection Level 4 is intended for less demanding applications (for example services addressed to fixed receivers).

2.2 Audio quality versus bit rate

The Eureka system uses a subset of the MUSICAM (ISO/IEC 11172-3 Layer II) audio standard. The available encoded bit rates range from 32 kbits/sec to 192 kbits/sec for a monophonic programme (or twice these bit rates for a stereophonic programme).

Four audio modes are supported:

The audio quality of a service depends on the bit rate used for the encoding of the audio signal. In general, the higher the bit rate the better the audio quality. Table 1-1 provides a rough guide of (current) data rates for a range of audio qualities.

ITU-R Recommendation BS1115 specifies use of ISO/IEC 11172-3 Layer II at 256 kbits/sec (stereo mode), for broadcast applications requiring compact disk (CD) quality. This recommendation is based on subjective listening tests undertaken in 1992. At the time, ISO/IEC 11172-3 Layer II at 192 kbits/sec (joint stereo mode) was also tested but was found to only marginally meet the audio quality requirement. Additional tests in 1993 failed to reveal sufficient improvement in the codec to warrant inclusion of this lower bit rate in the ITU recommendation.

Further listening tests were performed in 1995, as part of the US Electronic Industries Association’s (EIA) evaluation of DRB systems. A range of codecs were tested including ISO/IEC Layer II at 224 and 192 kbits/sec (joint stereo modes). The findings of this work indicate the Layer II codec at 224 kbits/sec is capable of meeting the basic audio quality criteria specified by the ITU-R. The lower rate of 192 kbits/sec again failed to meet the required quality.

It is worth noting that neither the ISO/IEC 1172-3 standard nor the Eureka 147 standard prescribe the audio encoder - only the format of the coded bit stream and actions to be taken by the decoder are specified. This approach has the advantage of allowing for future improvements in the encoder (i.e. further bit rate reductions) without the need to change existing decoders (i.e. receivers). The down-side, however, is that different encoders, although compliant with the ISO/IEC standard, may afford different audio qualities.

2.3 Programme Associated Data

Non-audio information can be carried within the encoded bit stream of each audio service. This information is referred to as programme associated data (PAD), and is generally reserved for data that has a strong relationship - in content and/or time - to the audio with which it is carried.

PAD is comprised of two parts. A fixed PAD (or F-PAD) is available by default and offers a net capacity of 667 bits/sec. This is available without prejudice to the audio data. An optional extended PAD (or X-PAD) can also be specified, providing an additional data capacity of up to 65 kbits/sec. In this case, however, the PAD is allocated from the available audio data capacity. The use of X-PAD may therefore result in a reduction in audio quality.

The following PAD applications have been specified:

(1) Dynamic range control (F-PAD): this feature can be used to improve the subjective audio quality in noisy environments, eg in some car listening situations.

(2) Music/speech indicator (F-PAD): this feature could allow a receiver to select specific sound processing based on audio content.

(3) ISRC and UPC/EAN (F-PAD): pre-recorded digital audio titles usually carry an International Standard Recording Code (ISRC) and Universal Product Code (UPC) or European Article Number (EAN). The ISRC is related to an individual commercial recording and is often included in the bit steam of each CD track. The UPC/EAN is a generic code used for a wide range of consumer products, and is normally indicated by a bar code on the back of the CD sleeve. Both, or either, of these codes can be conveyed in F-PAD to facilitate copyright accounting. The transmitted codes would be intended mainly for reception by special receivers.

(4) Command channel (F-PAD): this channel is designed to convey commands of a few bytes in length which require some degree of time relation to the audio. Although no applications have been defined yet, this facility could, for instance, be used to instruct the receiver to ‘switch-on’ a still picture accompanying music.

(5) In-house data (F-PAD): in-house data is intended only for use within the broadcasting organisation, using special receivers. Examples include signalling the source of a service, transmitter control and internal paging.

(6) Dynamic label (X-PAD): The dynamic label feature provides short textual messages, related to the current audio programme, for display on the receiver. This feature is comparable to radiotext in RDS.

(7) ITTS (X-PAD): The Interactive Test Transmission System (ITTS) provides a more comprehensive text facility from which the listener may select a topic of current interest. The text may describe the audio content, for example a song’s title or the name of the performer. The ITTS text repertoire may also contain information not directly related to the programme. For example, a stations name and logo, a programme listing, weather reports etc.

(8) Closed user-group packet & stream channels (X-PAD): These two channel types are designed to carry services in the X-PAD. Such services are not necessarily related to the audio programme, but remain under the control of the audio service provider. The content of the closed user-group channel is not identified, so specific ‘add-on’ devices are required to process the data. Some possible applications include data transfer, videotext and still pictures.

Because of the number of possibilities afforded by the Eureka system, the WorldDab Forum has issued a guideline on the functionalities and features (including PAD) expected of the first generation consumer receivers. The aim is to ensure availability and support - both by manufacturers and broadcasters - of a coherent (albeit limited) set of features, targeted initially at the car radio market. Amongst the PAD functions, only dynamic range control and dynamic label are identified as ‘essential’ while the music/speech indicator and X-PAD contents table (basically an index to the X-PAD) are listed as optional. Of course, the guideline does not prevent manufacturers from providing ‘high-end’ sets (e.g. for the home market) fitted with more elaborate features, but availability of many of these features to the consumer will also require the support of broadcasters.

3.1 Protection level

In addition to programme associated data which may be carried with the audio, general data may be conveyed as a separate service. This may take the form of a continuous stream or, alternatively, arranged in ‘packets’. In both cases the net capacity (i.e. before error protection) must be a multiple of 8 kbits/sec, though in the packet mode, data services of much lower capacity can be ‘bundled’ together (effectively forming a sub-multiplex) to satisfy this requirement.

As with audio services, the gross bit rate is dependent on the applied protection. For data services, four levels of protection (1..4) are specified. Level 4 is the lowest, and is equivalent to the lowest protection also afforded to audio services (i.e. Level 5). Level 1 provides maximum protection with a somewhat higher degree of redundancy than the highest protection level available to audio services.

3.2 Stream and Packet Mode

As noted in the previous section, two transport mechanisms are available for the transfer of data and these are referred to as either Stream or Packet mode.

Stream mode is suited to single data applications which can deliver a continuous stream at a constant bit rate. Since the net data rate must also correspond to multiples of 8 kbits/sec, the service provider must ensure information is available to ‘fill’ the sub-channel or, alternatively, employ techniques (such as bit stuffing), to make-up for any under utilization of the channel capacity by the application.

Data services carried in Stream mode may also be conveyed using Packet mode. In Packet mode the capacity of the sub-channel must also be a multiple of 8 kbits/sec, however the available capacity can be shared by more than one data service through use of addressable ‘packets’. This can increase transport efficiency when, for example, several service components have data rates well below the 8 kbits/sec breakpoints. There is a degree of trade-off, though. Part of the channel capacity is required for the organisation of the packet structure itself (headers, error checks etc) and this could amount to between 5 and 20 percent of the channel capacity. It is also evident that the use of Packet mode requires more sophisticated encoders and decoders. Principally for this last reason, it is unlikely that Packet mode will be supported in the first generation receivers. However, the feature will be required in the longer term if there is a significant amount of data broadcasting within the DRB system.

Table 1-2 lists the maximum number of audio sub-channels that can be accommodated in the MSC, along with the residual gross capacity, for a range of audio bit rates. This information has been tabulated for each of the five protection levels. It has been assumed that all audio services are encoded using the same protection level (which is likely to be the case in practice).

Examination of Table 1-2 shows that with a coding rate of 256 kbits/sec (the ITU recommended bit rate for CD quality) and Protection Level 3 (the minimum recommended for mobile reception), up to four services can be accommodated on the multiplex. If the same protection level is retained, the residual capacity of 256 kbits/sec could support an additional audio service at 128 kbits/sec (mono FM quality), or two 64 kbits/sec audio services (AM quality), or alternatively, one or more data services with an aggregate capacity of 128 kbits/sec.

However, in light of the recent EIA tests, it now seems reasonable to accept an audio bit rate of 224 kbits/sec for CD quality. On this basis, Table 1-2 shows that five services can be accommodated using Protection Level 3 . The residual capacity in this case is 64 kbits/sec, and is sufficient to support a voice grade audio channel at 32 kbits/sec or perhaps two data channels in Stream mode, each of 8 kbits/sec (again, this assumes Protection Level 3).

Return to 1997 Reports List or Return to Lab Home Page